Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Knowledge Unlearning on Internal e-commerce benchmark medium-scale seller 387 items (Forget Set)
Loading...
89.4
ROUGE
ME+GD
-3.576
20.562
44.7
68.838
May 9, 2025
ROUGE
Loss
Updated 4d ago
Evaluation Results
Method
Method
Links
ROUGE
Loss
ME+GD
Base Model=Llama 3.1 8B
2025.05
89.4
0.1
Baseline
Base Model=Llama 3.1 8B
2025.05
89.3
0.1
SimNPO+KL
Base Model=Llama 3.1 8...
2025.05
45.2
0.26
UnDIAL+KL
Base Model=Llama 3.1 8...
2025.05
44.3
0.17
RKLD+KL
Base Model=Llama 3.1 8B
2025.05
29.4
0.17
GA+KL
Base Model=Llama 3.1 8B
2025.05
20.4
0.47
UnDIAL+KL
Base Model=Llama 3.1 8...
2025.05
15.1
0.27
NPO+KL
Base Model=Llama 3.1 8...
2025.05
13.4
0.78
NPO
Base Model=Llama 3.1 8B
2025.05
13.1
1.06
GA
Base Model=Llama 3.1 8B
2025.05
11.6
5.03
NPO+KL
Base Model=Llama 3.1 8...
2025.05
11.5
0.89
SimNPO+KL
Base Model=Llama 3.1 8...
2025.05
0.3
33.63
Unilogit+KL
Base Model=Llama 3.1 8...
2025.05
0.2
6.64
Unilogit+KL
Base Model=Llama 3.1 8...
2025.05
0
10.78
Feedback
Search any
task
Search any
task